After changing local per interface beacon setting restore original
global beaconing state.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
When disable beaconing we clear register with beacon and newer set it
back, what make we stop send beacons infinitely.
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
syscore->resume() callback is expected to do not enable interrupts,
it generates warning like below otherwise:
[ 9386.365390] WARNING: CPU: 0 PID: 6733 at drivers/base/syscore.c:104 syscore_resume+0x9a/0xe0()
[ 9386.365403] Interrupts enabled after xen_acpi_processor_resume+0x0/0x34 [xen_acpi_processor]
...
[ 9386.365429] Call Trace:
[ 9386.365434] [<ffffffff81667a8b>] dump_stack+0x45/0x56
[ 9386.365437] [<ffffffff8106921d>] warn_slowpath_common+0x7d/0xa0
[ 9386.365439] [<ffffffff8106928c>] warn_slowpath_fmt+0x4c/0x50
[ 9386.365442] [<ffffffffa0261bb0>] ? xen_upload_processor_pm_data+0x300/0x300 [xen_acpi_processor]
[ 9386.365443] [<ffffffff814055fa>] syscore_resume+0x9a/0xe0
[ 9386.365445] [<ffffffff810aef42>] suspend_devices_and_enter+0x402/0x470
[ 9386.365447] [<ffffffff810af128>] pm_suspend+0x178/0x260
On xen_acpi_processor_resume() we call various procedures, which are
non atomic and can enable interrupts. To prevent the issue introduce
separate resume notify called after we enable interrupts on resume
and before we call other drivers resume callbacks.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Konrad Rzeszutek Wilk <[email protected]>
This reverts commit eac40d9631. It cause
random connection drops on RT5390 PCI adapters.
On Mediatek there is different driver version available for RT53xx chip
based on bus type (2.5.0.3 for PCI and 2.6.1.3 for USB). Hence possibly
we should set registers differently based on bus type. But is also
possible that new driver (i.e. 2.6.1.3) was not verified on RT53xx USB.
Until we figure out how to initialize registers properly for RT53xx just
revert commit eac40d9631 since it cause
regression.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
iwlegacy version of this iwlwifi patch:
commit aed7d9ac18
Author: Emmanuel Grumbach <[email protected]>
Date: Wed Feb 20 11:33:00 2013 +0200
iwlwifi: disable 8K A-MSDU by default
Supporting 8K A-MSDU means that we need to allocate order 1
pages for every Rx packet. Even when there is no traffic.
This adds stress on the memory manager. The handling of
compound pages is also less trivial for the memory manager
and not using them will make the allocation code run faster
although I didn't really measure.
Eric also pointed out that having huge buffers with little
data in them is not very nice towards the TCP stack since
the truesize of the skb is huge. This doesn't allow TCP
to have a big Rx window.
See https://patchwork.kernel.org/patch/2167711/ for details.
Note that very few vendors will actually send A-MSDU.
Disable this feature by default.
Signed-off-by: Emmanuel Grumbach <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
ehci_irq() and ehci_hrtimer_func() can deadlock on ehci->lock when
threadirqs option is used. To prevent the deadlock use
spin_lock_irqsave() in ehci_irq().
This change can be reverted when hrtimer callbacks become threaded.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Cc: stable <[email protected]>
Acked-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Even if we mark PS on, device still worked in normal mode. Patch
corrects that and now we send proper powertable command to device,
which put it in sleep mode when PS is on.
Reported-and-tested-by: Tino Keitel <[email protected]>
Tested-by: Pedro Francisco <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Merge reclaim check for 3945 & 4965. This add some more checks for
3945, most importantly N_RX notify.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
We have only 5 tx queues on 3945, updating il->txq[5] results in
writing random value to HBUS_TARG_WRPTR register.
Additionally use spin lock to protect txq->write_ptr and
txq->need_update fields also modified in TX path.
Tested-by: Pedro Francisco <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
On commit 28f2bce9f8 I make change that
print various messages as default. This can cause flood of messages
related to TX status timeout on some environments. I partially fixed
problem on commit bb9c298f31, but forgot
to move two more messages to debug level.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
mac80211 ->sta_rc_update() callback must be atomic. Since we have to
take mutex and do other operations that can sleep when sending fimrware
commands to device, the only option to satisfy atomicity requirement of
->sta_rc_update(), that I can see, is introduce work_struct and defer
uploading new rates to that work.
Tested-by: Oleksij Rempel <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
It is know that PS cause issues on that old devices, disable it by
default.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
We have disabled it currently on other buses. PS can cause some issues,
not necessarily with our driver but on AP, that are not easy to debug.
Since behaviour differs on rt2800usb and rt2800pci, user usually blame
for malfunction rt2800usb driver, whereas issue is on AP side.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
sta_rc_update() callback must be atomic, hence we can not take mutexes
or do other operations, which can sleep in ath9k_htc_sta_rc_update().
I think we can just return from ath9k_htc_sta_rc_update(), if it is
called without IEEE80211_RC_SUPP_RATES_CHANGED bit. That will help
with scheduling while atomic bug for most cases (except mesh and IBSS
modes).
For mesh and IBSS I do not see other solution like creating additional
workqueue, because sending firmware command require us to sleep, but
this can be done in additional patch.
Patch partially fixes bug:
https://bugzilla.redhat.com/show_bug.cgi?id=990955
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Not removing pm qos request and free memory for it can cause crash,
when some other driver use pm qos. For example, this oops:
BUG: unable to handle kernel paging request at fffffffffffffff8
IP: [<ffffffff81307a6b>] plist_add+0x5b/0xd0
Call Trace:
[<ffffffff810acf25>] pm_qos_update_target+0x125/0x1e0
[<ffffffff810ad071>] pm_qos_add_request+0x91/0x100
[<ffffffffa053ec14>] e1000_open+0xe4/0x5b0 [e1000e]
was caused by earlier i915 probe failure:
[drm:i915_report_and_clear_eir] *ERROR* EIR stuck: 0x00000010, masking
[drm:init_ring_common] *ERROR* render ring initialization failed ctl 0001f001 head 00003004 tail 00000000 start 00003000
[drm:i915_driver_load] *ERROR* failed to init modeset
i915: probe of 0000:00:02.0 failed with error -5
Bug report:
http://bugzilla.redhat.com/show_bug.cgi?id=1057533
Reported-by: Giandomenico De Tullio <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
[danvet: Drop unnecessary code movement.]
Signed-off-by: Daniel Vetter <[email protected]>
We can mess logs if user space try to open device again and again if
RFKILL switch is on. Do not print message and return ERFKILL error
instead to indicate where the problem is.
Note that iwl4965 handle this problem differently, it allows to open
device when radio is disabled.
Reported-by: Dietmar Rudolph <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Since:
commit 36323f817a
Author: Thomas Huehn <[email protected]>
Date: Mon Jul 23 21:33:42 2012 +0200
mac80211: move TX station pointer and restructure TX
we do not pass sta pointer to rt2x00queue_create_tx_descriptor_ht(),
hence we do not correctly set station WCID and AMPDU density parameters.
Cc: [email protected] # 3.7+
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Polling TX statuses too frequently has two negative effects. First is
randomly peek CPU usage, causing overall system functioning delays.
Second bad effect is that device is not able to fill TX statuses in
H/W register on some workloads and we get lot of timeouts like below:
ieee80211 phy4: rt2800usb_entry_txstatus_timeout: Warning - TX status timeout for entry 7 in queue 2
ieee80211 phy4: rt2800usb_entry_txstatus_timeout: Warning - TX status timeout for entry 7 in queue 2
ieee80211 phy4: rt2800usb_txdone: Warning - Got TX status for an empty queue 2, dropping
This not only cause flood of messages in dmesg, but also bad throughput,
since rate scaling algorithm can not work optimally.
In the future, we should probably make polling interval be adjusted
automatically, but for now just increase values, this make mentioned
problems gone.
Resolve:
https://bugzilla.kernel.org/show_bug.cgi?id=62781
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
When we restart firmware and it is marked as not alive, we can still get
calls from mac80211. Don't WARN on in this situation as this triggers
automatic bug reports with no valuable information.
This continuation of:
commit 8ca95995e6
Author: Emmanuel Grumbach <[email protected]>
Date: Sun Sep 15 11:37:17 2013 +0300
iwlwifi: don't WARN on host commands sent when firmware is dead
which remove WARN_ONCE from one place, but those warnings are also
triggered from other functions.
Patch also adds unlikely() statement.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Emmanuel Grumbach <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Add support for new RF chip ID: 3070. It seems to be the same as 5370,
maybe vendor just put wrong value on the eeprom, but add this id anyway
since devices with it showed on the marked.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Since when we detect beacon lost we do active AP probing (using nullfunc
frame or probe request) there is no need to have beacon polling. Flags
IEEE80211_STA_BEACON_POLL seems to be used just for historical reasons.
Change also make that after we start connection poll due to beacon loss,
next received beacon will abort the poll.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
My commit
commit c630ccf1a1
Author: Stanislaw Gruszka <[email protected]>
Date: Sat Mar 16 19:19:46 2013 +0100
rt2800: rearrange bbp/rfcsr initialization
make Maxim machine freeze when try to start wireless device.
Initialization order and sending MCU_BOOT_SIGNAL request, changed in
above commit, is important. Doing things incorrectly make PCIe bus
problems, which can froze the machine.
This patch change initialization sequence like vendor driver do:
function NICInitializeAsic() from
2011_1007_RT5390_RT5392_Linux_STA_V2.5.0.3_DPO (PCI devices) and
DPO_RT5572_LinuxSTA_2.6.1.3_20121022 (according Mediatek, latest driver
for RT8070/RT3070/RT3370/RT3572/RT5370/RT5372/RT5572 USB devices).
It fixes freezes on Maxim system.
Resolve:
https://bugzilla.redhat.com/show_bug.cgi?id=1000679
Reported-and-tested-by: Maxim Polyakov <[email protected]>
Bisected-by: Igor Gnatenko <[email protected]>
Cc: [email protected] # 3.10+
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
scale_stime() silently assumes that stime < rtime, otherwise
when stime == rtime and both values are big enough (operations
on them do not fit in 32 bits), the resulting scaling stime can
be bigger than rtime. In consequence utime = rtime - stime
results in negative value.
User space visible symptoms of the bug are overflowed TIME
values on ps/top, for example:
$ ps aux | grep rcu
root 8 0.0 0.0 0 0 ? S 12:42 0:00 [rcuc/0]
root 9 0.0 0.0 0 0 ? S 12:42 0:00 [rcub/0]
root 10 62422329 0.0 0 0 ? R 12:42 21114581:37 [rcu_preempt]
root 11 0.1 0.0 0 0 ? S 12:42 0:02 [rcuop/0]
root 12 62422329 0.0 0 0 ? S 12:42 21114581:35 [rcuop/1]
root 10 62422329 0.0 0 0 ? R 12:42 21114581:37 [rcu_preempt]
or overflowed utime values read directly from /proc/$PID/stat
Reference:
https://lkml.org/lkml/2013/8/20/259
Reported-and-tested-by: Sergey Senozhatsky <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Cc: [email protected]
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul E. McKenney <[email protected]>
Cc: Borislav Petkov <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
We should not do temperature compensation on devices without
EXTERNAL_TX_ALC bit set (called DynamicTxAgcControl on vendor driver).
Such devices can have totally bogus TSSI parameters on the EEPROM,
but still threaded by us as valid and result doing wrong TX power
calculations.
This fix inability to connect to AP on slightly longer distance on
some Ralink chips/devices.
Reported-and-tested-by: Fabien ADAM <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
My current 3.11 fix:
commit 788f7a56fc
Author: Stanislaw Gruszka <[email protected]>
Date: Thu Aug 1 12:07:55 2013 +0200
iwl4965: reset firmware after rfkill off
broke rfkill notification to user-space . I missed that bug, because
I compiled without CONFIG_RFKILL, sorry about that.
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
If device was put into a sleep and system was restarted or module
reloaded, we have to wake device up before sending other commands.
Otherwise it will fail to start with Microcode error.
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
4965 version of Eric patch "iwl3945: better skb management in rx path".
It fixes several problems :
1) skb->truesize is underestimated.
We really consume PAGE_SIZE bytes for a fragment,
not the frame length.
2) 128 bytes of initial headroom is a bit low and forces reallocations.
3) We can avoid consuming a full page for small enough frames.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Eric Dumazet <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
If channel switch is pending and we remove interface we can
crash like showed below due to passing NULL vif to mac80211:
BUG: unable to handle kernel paging request at fffffffffffff8cc
IP: [<ffffffff8130924d>] strnlen+0xd/0x40
Call Trace:
[<ffffffff8130ad2e>] string.isra.3+0x3e/0xd0
[<ffffffff8130bf99>] vsnprintf+0x219/0x640
[<ffffffff8130c481>] vscnprintf+0x11/0x30
[<ffffffff81061585>] vprintk_emit+0x115/0x4f0
[<ffffffff81657bd5>] printk+0x61/0x63
[<ffffffffa048987f>] ieee80211_chswitch_done+0xaf/0xd0 [mac80211]
[<ffffffffa04e7b34>] iwl_chswitch_done+0x34/0x40 [iwldvm]
[<ffffffffa04f83c3>] iwlagn_commit_rxon+0x2a3/0xdc0 [iwldvm]
[<ffffffffa04ebc50>] ? iwlagn_set_rxon_chain+0x180/0x2c0 [iwldvm]
[<ffffffffa04e5e76>] iwl_set_mode+0x36/0x40 [iwldvm]
[<ffffffffa04e5f0d>] iwlagn_mac_remove_interface+0x8d/0x1b0 [iwldvm]
[<ffffffffa0459b3d>] ieee80211_do_stop+0x29d/0x7f0 [mac80211]
This is because we nulify ctx->vif in iwlagn_mac_remove_interface()
before calling some other functions that teardown interface. To fix
just check ctx->vif on iwl_chswitch_done(). We should not call
ieee80211_chswitch_done() as channel switch works were already canceled
by mac80211 in ieee80211_do_stop() -> ieee80211_mgd_stop().
Resolve:
https://bugzilla.redhat.com/show_bug.cgi?id=979581
Cc: [email protected]
Reported-by: Lukasz Jagiello <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Reviewed-by: Emmanuel Grumbach <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
This reverts commit dd9c46408f.
With "iwl{4965,3495): workaround for firmware frame tx rejection"
patches we can enable IEEE80211_HW_REPORTS_TX_ACK_STATUS again.
Tested-by: Jake Edge <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Firmware can reject to transmit frame on passive channel, when it
did not yet received any frame with valid CRC on that channel.
Workaround this problem in the driver.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Firmware can reject to transmit frame on passive channel, when it
did not yet received any frame with valid CRC on that channel.
Workaround this problem in the driver.
Tested-by: Jake Edge <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Add functions that will stop/wake all queues. Make them safe
regarding multiple calls and when some ac are stopped/woke
independently.
Tested-by: Jake Edge <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
My change:
commit cee2c7315f
Author: Stanislaw Gruszka <[email protected]>
Date: Fri Oct 5 13:44:09 2012 +0200
rt2800: use BBP_R1 for setting tx power
unfortunately does not work well with RT5390 and RT3290 chips as they
require different temperature compensation TX power settings (TSSI
tuning). Since that commit make wireless connection very unstable on
those chips, restore previous behavior to fix regression. Once we
implement proper TSSI tuning on 5390/3290 we can restore back setting
TX power by BBP_R1 register for those chips.
Reported-and-tested-by: Mike Romberg <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Since driver does not use control.rates[0].count, we have never set that
variable. But currently, after rate control API rewrite, this is required
by mac80211. Otherwise legacy rates control does not work and we transmit
always at 1Mbit/s on pre 11n networks.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
We send direct probe to broadcast address, as some APs do not respond to
unicast PROBE frames when unassociated. Broadcast frames are not acked,
so we can not use that for trigger MLME state machine, but we need to
use old timeout mechanism.
This fixes authentication timed out like below:
[ 1024.671974] wlan6: authenticate with 54:e6:fc:98:63:fe
[ 1024.694125] wlan6: direct probe to 54:e6:fc:98:63:fe (try 1/3)
[ 1024.695450] wlan6: direct probe to 54:e6:fc:98:63:fe (try 2/3)
[ 1024.700586] wlan6: send auth to 54:e6:fc:98:63:fe (try 3/3)
[ 1024.701441] wlan6: authentication with 54:e6:fc:98:63:fe timed out
With fix, we have:
[ 4524.198978] wlan6: authenticate with 54:e6:fc:98:63:fe
[ 4524.220692] wlan6: direct probe to 54:e6:fc:98:63:fe (try 1/3)
[ 4524.421784] wlan6: send auth to 54:e6:fc:98:63:fe (try 2/3)
[ 4524.423272] wlan6: authenticated
[ 4524.423811] wlan6: associate with 54:e6:fc:98:63:fe (try 1/3)
[ 4524.427492] wlan6: RX AssocResp from 54:e6:fc:98:63:fe (capab=0x431 status=0 aid=1)
Cc: [email protected] # 3.9
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Jake reported that since commit 1672c0e319
"mac80211: start auth/assoc timeout on frame status", he is unable to
connect to his AP, which is configured to use passive channel.
After switch to passive channel 4965 firmware drops any TX packet until
it receives beacon. Before commit 1672c0e3 we waited on channel and
retransmit packet after 200ms, that makes we receive beacon on the
meantime and association process succeed. New mac80211 behaviour cause
that any ASSOC frame fail immediately on iwl4965 and we can not
associate.
This patch restore old mac80211 behaviour for iwl4965, by removing
IEEE80211_HW_REPORTS_TX_ACK_STATUS feature. This feature will be
added again to iwl4965 driver, when different, more complex
workaround for this firmware issue, will be added to the driver.
Cc: [email protected] # 3.9
Bisected-by: Jake Edge <[email protected]>
Reported-and-tested-by: Jake Edge <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
This patch claim to fix "WARNING: at net/mac80211/util.c:599
ieee80211_can_queue_work.isra.7+0x30/0x40", which was reported at:
https://bugzilla.redhat.com/show_bug.cgi?id=922295
We use ATH_STAT_STARTED flag to disallow to perform
ath5k_tx_complete_poll_work() code, hence reschedule
ah->tx_complete_work, when we stop device. This flag was defined in
ath5k code, but it was not used.
I didn't get feedback if the fix works, so patch is compile only tested.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Dave Hansen reported strange utime/stime values on his system:
https://lkml.org/lkml/2013/4/4/435
This happens because prev->stime value is bigger than rtime
value. Root of the problem are non-monotonic rtime values (i.e.
current rtime is smaller than previous rtime) and that should be
debugged and fixed.
But since problem did not manifest itself before commit
62188451f0 "cputime: Avoid
multiplication overflow on utime scaling", it should be threated
as regression, which we can easily fixed on cputime_adjust()
function.
For now, let's apply this fix, but further work is needed to fix
root of the problem.
Reported-and-tested-by: Dave Hansen <[email protected]>
Cc: <[email protected]> # 3.9+
Signed-off-by: Stanislaw Gruszka <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: [email protected]
Cc: Linus Torvalds <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
Due to rounding in scale_stime(), for big numbers, scaled stime
values will grow in chunks. Since rtime grow in jiffies and we
calculate utime like below:
prev->stime = max(prev->stime, stime);
prev->utime = max(prev->utime, rtime - prev->stime);
we could erroneously account stime values as utime. To prevent
that only update prev->{u,s}time values when they are smaller
than current rtime.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: [email protected]
Cc: Linus Torvalds <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
Here is patch, which adds Linus's cputime scaling algorithm to the
kernel.
This is a follow up (well, fix) to commit
d9a3c9823a ("sched: Lower chances
of cputime scaling overflow") which commit tried to avoid
multiplication overflow, but did not guarantee that the overflow
would not happen.
Linus crated a different algorithm, which completely avoids the
multiplication overflow by dropping precision when numbers are
big.
It was tested by me and it gives good relative error of
scaled numbers. Testing method is described here:
http://marc.info/?l=linux-kernel&m=136733059505406&w=2
Originally-From: Linus Torvalds <[email protected]>
Signed-off-by: Stanislaw Gruszka <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: [email protected]
Cc: Dave Hansen <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
We can check buf against NULL instead of having additional bool
variable.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Reviewed-by: Emmanuel Grumbach <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
5ghz_disable has no effect any longer, that was changed during
refactoring of EEPROM reading/parsing. Remove it, wpa_supplicant
allow now to specify frequencies, on which device will operate.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Reviewed-by: Emmanuel Grumbach <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
This basically reverts commit b207cdb07f.
Now is possible to use drv_{add,remove}_interface() and vif->debugfs_dir
to create/remove per interface debugfs files. Remove redundant
callbacks.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
There is need create driver own per interface debugfs files. This is
currently done by drv_{add,remove}_interface_debugfs() callbacks. But it
is possible that after we remove interface from the driver (i.e.
on suspend) we call drv_remove_interface_debugfs() function. Fixing this
problem will require to add call drv_{add,remove}_interface_debugfs()
anytime we create and remove interface in mac80211. So it's better to
add debugfs dir dentry to vif structure to allow to create/remove
custom debugfs driver files on drv_{add,remove}_interface().
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Since now we disconnect before suspend, various code which save
connection state can now be removed from suspend and resume
procedure. Cleanup on resume side is smaller as ieee80211_reconfig()
is also used for H/W restart.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
If possible that after suspend, cfg80211 will receive request to
disconnect what require action on interface that was removed during
suspend.
Problem can manifest itself by various warnings similar to below one:
WARNING: at net/mac80211/driver-ops.h:12 ieee80211_bss_info_change_notify+0x2f9/0x300 [mac80211]()
wlan0: Failed check-sdata-in-driver check, flags: 0x4
Call Trace:
[<c043e0b3>] warn_slowpath_fmt+0x33/0x40
[<f83707c9>] ieee80211_bss_info_change_notify+0x2f9/0x300 [mac80211]
[<f83a660a>] ieee80211_recalc_ps_vif+0x2a/0x30 [mac80211]
[<f83a6706>] ieee80211_set_disassoc+0xf6/0x500 [mac80211]
[<f83a9441>] ieee80211_mgd_deauth+0x1f1/0x280 [mac80211]
[<f8381b36>] ieee80211_deauth+0x16/0x20 [mac80211]
[<f8261e70>] cfg80211_mlme_down+0x70/0xc0 [cfg80211]
[<f8264de1>] __cfg80211_disconnect+0x1b1/0x1d0 [cfg80211]
To fix the problem disconnect from any associated network before
suspend. User space is responsible to establish connection again
after resume. This basically need to be done by user space anyway,
because associated stations can go away during suspend (for example
NetworkManager disconnects on suspend and connect on resume by default).
Patch also handle situation when driver refuse to suspend with wowlan
configured and try to suspend again without it.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Since two years no mac80211 driver implement support for NAPI. Looks
this feature is unneeded, so remove it from generic mac80211 code.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
This patch check output of pci_map_single() calls. I missed them on
my previous patch "iwlegacy: check for dma mapping errors", which
fixed only pci_map_page() calls.
To handle remaining possible dma mappings errors, we need to rearrange
ilXXXX_tx_skb() and il_enqueue_hcmd() functions.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Check output of dma_map_single functions which nowadays can fail (when
IOMMU is used). On write_beacon callbacks just print error, similar
like padding error is handled by rt2800_write_beacon.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Patch vastly improve latency while scanning. Slight throughput
improvements were observed as well. Is intended for improve performance
of voice and video applications, when scan is periodically requested by
user space (i.e. default NetworkManager behaviour).
Patch remove latency requirement based on PM_QOS_NETWORK_LATENCY,
this value is 2000 seconds by default (i.e. approximately 0.5 hour !?!).
Also remove listen interval requirement, which based on beaconing and
depending on BSS parameters. It can make we stay off-channel for a
second or more.
Instead try to offer the best latency that we could, i.e. be off-channel
no longer than PASSIVE channel scan time: 125 ms. That mean we will
scan two ACTIVE channels and go back to on-channel, and one PASSIVE
channel, and go back to on-channel.
Patch also decrease PASSIVE channel scan time to about 110 ms.
As drawback patch increase overall scan time. On my tests, when scanning
both 2GHz and 5GHz bands, scanning time increase from 5 seconds up to 10
seconds. Since that increase happen only when we are associated, I think
it can be acceptable. If eventually better scan time is needed for
situations when we lose signal and quickly need to decide to which AP
roam, additional scan flag or parameter can be introduced.
I tested patch by doing:
while true; do iw dev wlan0 scan; sleep 3; done > /dev/null
and
ping -i0.2 -c 1000 HOST
on remote and local machine, results are as below:
* Ping from local periodically scanning machine to AP:
Unpatched: rtt min/avg/max/mdev = 0.928/24.946/182.135/36.873 ms
Patched: rtt min/avg/max/mdev = 0.928/19.678/150.845/33.130 ms
* Ping from remote machine to periodically scanning machine:
Unpatched: rtt min/avg/max/mdev = 1.637/120.683/709.139/164.337 ms
Patched: rtt min/avg/max/mdev = 1.807/26.893/201.435/40.284 ms
Throughput measured by scp show following results.
* Upload to periodically scanning machine:
Unpatched: 3.9MB/s 03:15
Patched: 4.3MB/s 02:58
* Download from periodically scanning machine:
Unpatched: 5.5MB/s 02:17
Patched: 6.2MB/s 02:02
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
This is basically a revert of:
commit 5b632fe85e
Author: Stanislaw Gruszka <[email protected]>
Date: Mon Dec 3 12:56:33 2012 +0100
mac80211: introduce IEEE80211_HW_TEARDOWN_AGGR_ON_BAR_FAIL
We do not need this flag any longer, rt2x00 BAR/BA problem was fixed
correctly by wireless-testing commit:
commit 84e9e8ebd3
Author: Helmut Schaa <[email protected]>
Date: Thu Jan 17 17:34:32 2013 +0100
rt2x00: Improve TX status handling for BlockAckReq frames
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Since:
commit b23b025fe2
Author: Ben Greear <[email protected]>
Date: Fri Feb 4 11:54:17 2011 -0800
mac80211: Optimize scans on current operating channel.
we do not disable PS while going back to operational channel (on
ieee80211_scan_state_suspend) and deffer that until scan finish.
But since we are allowed to send frames, we can send a frame to AP
without PM bit set, so disable PS on AP side. Then when we switch
to off-channel (in ieee80211_scan_state_resume) we do not enable PS.
Hence we are off-channel with PS disabled, frames are not buffered
by AP.
To fix remove offchannel_ps_disable argument and always enable PS when
going off-channel and disable it when going on-channel, like it was
before.
Cc: [email protected] # 2.6.39+
Signed-off-by: Stanislaw Gruszka <[email protected]>
Tested-by: Seth Forshee <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Since:
commit 2c60db0370
Author: Eric Dumazet <[email protected]>
Date: Sun Sep 16 09:17:26 2012 +0000
net: provide a default dev->ethtool_ops
wireless core does not correctly assign ethtool_ops.
After alloc_netdev*() call, some cfg80211 drivers provide they own
ethtool_ops, but some do not. For them, wireless core provide generic
cfg80211_ethtool_ops, which is assigned in NETDEV_REGISTER notify call:
if (!dev->ethtool_ops)
dev->ethtool_ops = &cfg80211_ethtool_ops;
But after Eric's commit, dev->ethtool_ops is no longer NULL (on cfg80211
drivers without custom ethtool_ops), but points to &default_ethtool_ops.
In order to fix the problem, provide function which will overwrite
default_ethtool_ops and use it by wireless core.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Johannes Berg <[email protected]>
Acked-by: Ben Hutchings <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Power save support was removed from iwlegacy due to possible firmware
crashes problems it cause. I use to plan first inspect code to find
reason of problems, fix them and then allow to enable PS. But
realistically - code inspection will not happen, so let's do it, and
wait for eventual bug reports.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Dump implementation of flush, which just wait until all TX queues
become empty.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
This revert:
commit be03d4a45c
Author: Andreas Hartmann <[email protected]>
Date: Tue Apr 17 00:25:28 2012 +0200
rt2x00: Don't let mac80211 send a BAR when an AMPDU subframe fails
To fix problem workaround by above commit use
IEEE80211_HW_TEARDOWN_AGGR_ON_BAR_FAIL flag (see change log for
"mac80211: introduce IEEE80211_HW_TEARDOWN_AGGR_ON_BAR_FAIL" patch).
Resolve: https://bugzilla.kernel.org/show_bug.cgi?id=42828
Bisected-by: Francisco Pina Martins <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Commit f0425beda4 "mac80211: retry sending
failed BAR frames later instead of tearing down aggr" caused regression
on rt2x00 hardware (connection hangs). This regression was fixed by
commit be03d4a45c "rt2x00: Don't let
mac80211 send a BAR when an AMPDU subframe fails". But the latter
commit caused yet another problem reported in
https://bugzilla.kernel.org/show_bug.cgi?id=42828#c22
After long discussion in this thread:
http://mid.gmane.org/[email protected]
and testing various alternative solutions, which failed on one or other
setup, we have no other good fix for the issues like just revert both
mentioned earlier commits.
To do not affect other hardware which benefit from commit
f0425beda4, instead of reverting it,
introduce flag that when used will restore mac80211 behaviour before
the commit.
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
[replaced link with mid.gmane.org that has message-id]
Signed-off-by: Johannes Berg <[email protected]>
Some hardware has correct (!= 0xff) value of tssi_bounds[4] in the
EEPROM, but step is equal to 0xff. This results on ridiculous delta
calculations and completely broke TX power settings.
Reported-and-tested-by: Pavel Lucik <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Patch fixes warnings like below happened on resume:
WARNING: at net/mac80211/driver-ops.h:12 check_sdata_in_driver+0x32/0x34()
Problem is that in __ieee80211_susped() we remove sdata (i.e wlan0
interface) and then during resume we call usb_unbind_interface() ->
ieee80211_unregister_hw() with sdata removed.
Patch fixes problem by adding .reset_resume calback, hence we do not
unbind usb device on resume. This callback can be the same as normal
.resume callback, sice we do all needed initalization during interface
start, which is performed on resume [ ieee80211_resume() ->
ieee80211_reconfig() -> rt2x00mac_start() -> rt2x00lib_start ].
Resolves:
https://bugzilla.kernel.org/show_bug.cgi?id=48041
Reported-by: David Herrmann <[email protected]>
Reported-and-tested-by: Stephen Boyd <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Some rt2800 devices don't have their calibrated max eirp tx power in
their calibration data. For those devices reduce tx power according to
difference between regulatory max channel power and requested tx power.
This patch is based on Helmut Schaa work.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Acked-by: Helmut Schaa <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Preparation for use regulatory max channel power in TX power delta
calculations.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Don use TX_PWR_CFG_0 register value of OFDM 6M tx power as criterion
since it can be changed. The same do vendor driver (see
AsicAdjustSingleSkuTxPower and AsicGetTxPowerOffset functions from
2011_0719_RT3070_RT3370_RT5370_RT5372_Linux_STA_V2.5.0.3_DPO).
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
We skip compensate calculation for non 11b rates on 2.4GHz band. I do
not see that on vendor driver
(2011_0719_RT3070_RT3370_RT5370_RT5372_Linux_STA_V2.5.0.3_DPO).
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Helmut Schaa <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Based on AsicAdjustTxPower function from vendor driver
(2011_0719_RT3070_RT3370_RT5370_RT5372_Linux_STA_V2.5.0.3_DPO)
limit per rate TX power values we program into TX_PWR_CFG_ registers.
Note that on some configurations (devices/rates) is allowed to use
bigger values than 0xc, but we use safe maximum value for now. Further
work need to be done to allow use bigger values than 0xc.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
TX power delta can be negative. TX_PWR_CFG_ registers allow to set delta
only in range between 0 dBm and 15 dBm (4 bits for each rate). Se we
need to use BBP_R1 to configure negative deltas.
Not utilize +6 dBm increasing BBP_R1 option for safety reason. For now,
this can be used for devices, which export maximum allowed TX power
value.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Helmut Schaa <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Acked-by: Ivo van Doorn <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
This patch fix corruption which can manifest itself by following crash
when switching on rfkill switch with rt2x00 driver:
https://bugzilla.redhat.com/attachment.cgi?id=615362
Pointer key->u.ccmp.tfm of group key get corrupted in:
ieee80211_rx_h_michael_mic_verify():
/* update IV in key information to be able to detect replays */
rx->key->u.tkip.rx[rx->security_idx].iv32 = rx->tkip_iv32;
rx->key->u.tkip.rx[rx->security_idx].iv16 = rx->tkip_iv16;
because rt2x00 always set RX_FLAG_MMIC_STRIPPED, even if key is not TKIP.
We already check type of the key in different path in
ieee80211_rx_h_michael_mic_verify() function, so adding additional
check here is reasonable.
Cc: [email protected] # 3.0+
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
On architectures where cputime_t is 64 bit type, is possible to trigger
divide by zero on do_div(temp, (__force u32) total) line, if total is a
non zero number but has lower 32 bit's zeroed. Removing casting is not
a good solution since some do_div() implementations do cast to u32
internally.
This problem can be triggered in practice on very long lived processes:
PID: 2331 TASK: ffff880472814b00 CPU: 2 COMMAND: "oraagent.bin"
#0 [ffff880472a51b70] machine_kexec at ffffffff8103214b
#1 [ffff880472a51bd0] crash_kexec at ffffffff810b91c2
#2 [ffff880472a51ca0] oops_end at ffffffff814f0b00
#3 [ffff880472a51cd0] die at ffffffff8100f26b
#4 [ffff880472a51d00] do_trap at ffffffff814f03f4
#5 [ffff880472a51d60] do_divide_error at ffffffff8100cfff
#6 [ffff880472a51e00] divide_error at ffffffff8100be7b
[exception RIP: thread_group_times+0x56]
RIP: ffffffff81056a16 RSP: ffff880472a51eb8 RFLAGS: 00010046
RAX: bc3572c9fe12d194 RBX: ffff880874150800 RCX: 0000000110266fad
RDX: 0000000000000000 RSI: ffff880472a51eb8 RDI: 001038ae7d9633dc
RBP: ffff880472a51ef8 R8: 00000000b10a3a64 R9: ffff880874150800
R10: 00007fcba27ab680 R11: 0000000000000202 R12: ffff880472a51f08
R13: ffff880472a51f10 R14: 0000000000000000 R15: 0000000000000007
ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018
#7 [ffff880472a51f00] do_sys_times at ffffffff8108845d
#8 [ffff880472a51f40] sys_times at ffffffff81088524
#9 [ffff880472a51f80] system_call_fastpath at ffffffff8100b0f2
RIP: 0000003808caac3a RSP: 00007fcba27ab6d8 RFLAGS: 00000202
RAX: 0000000000000064 RBX: ffffffff8100b0f2 RCX: 0000000000000000
RDX: 00007fcba27ab6e0 RSI: 000000000076d58e RDI: 00007fcba27ab6e0
RBP: 00007fcba27ab700 R8: 0000000000000020 R9: 000000000000091b
R10: 00007fcba27ab680 R11: 0000000000000202 R12: 00007fff9ca41940
R13: 0000000000000000 R14: 00007fcba27ac9c0 R15: 00007fff9ca41940
ORIG_RAX: 0000000000000064 CS: 0033 SS: 002b
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: Peter Zijlstra <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Thomas Gleixner <[email protected]>
commit eccc068e8e
Author: Hong Wu <[email protected]>
Date: Wed Jan 11 20:33:39 2012 +0200
wireless: Save original maximum regulatory transmission power for the calucation of the local maximum transmit pow
changed the way we calculate chan->max_power as min(chan->max_power,
chan->max_reg_power). That broke rt2x00 (and perhaps some other
drivers) that do not set chan->max_power. It is not so easy to fix this
problem correctly in rt2x00.
According to commit eccc068e8 changelog, change claim only to save
maximum regulatory power - changing setting of chan->max_power was side
effect. This patch restore previous calculations of chan->max_power and
do not touch chan->max_reg_power.
Cc: [email protected] # 3.4+
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Luis R. Rodriguez <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Data pointer on rt2x00queue_for_each_entry() is never used - remove it.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Merge 3290 and 5390 POWER_BOUND and FREQ_OFFSET_BOUND defines.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Helmut Schaa <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Fix indention and remove unnecessary brackets and compares.
Signed-off-by: Stanislaw Gruszka <[email protected]>
Acked-by: Helmut Schaa <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
This is iwlegacy version of:
commit 342bbf3fee
Author: Johannes Berg <[email protected]>
Date: Sun Mar 4 08:50:46 2012 -0800
iwlwifi: always monitor for stuck queues
If we only monitor while associated, the following
can happen:
- we're associated, and the queue stuck check
runs, setting the queue "touch" time to X
- we disassociate, stopping the monitoring,
which leaves the time set to X
- almost 2s later, we associate, and enqueue
a frame
- before the frame is transmitted, we monitor
for stuck queues, and find the time set to
X, although it is now later than X + 2000ms,
so we decide that the queue is stuck and
erroneously restart the device
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
On rt2x00_dmastart() we increase index specified by Q_INDEX and on
rt2x00_dmadone() we increase index specified by Q_INDEX_DONE. So entries
between Q_INDEX_DONE and Q_INDEX are those we currently process in the
hardware. Entries between Q_INDEX and Q_INDEX_DONE are those we can
submit to the hardware.
According to that fix rt2x00usb_kick_queue(), as we need to submit RX
entries that are not processed by the hardware. It worked before only
for empty queue, otherwise was broken.
Note that for TX queues indexes ordering are ok. We need to kick entries
that have filled skb, but was not submitted to the hardware, i.e.
started from Q_INDEX_DONE and have ENTRY_DATA_PENDING bit set.
From practical standpoint this fixes RX queue stall, usually reproducible
in AP mode, like for example reported here:
https://bugzilla.redhat.com/show_bug.cgi?id=828824
Reported-and-tested-by: Franco Miceli <[email protected]>
Reported-and-tested-by: Tom Horsley <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Stanislaw Gruszka